We report the first comparative evaluation between the Bruker Biotyper MS (BMS) and the Vitek MS (VMS) for the identification of yeasts. The rate of correct identifications at the species level was comparable using the commercial databases (89.8% versus 84.3%; P ؍ 0.712), but higher for BMS using an in-house-extended database (100% versus 84.3%; P ؍ 0.245). Importantly, the rate of misidentification was significantly higher for VMS (1% versus 12.1%; P < 0.0001), including the rate of major errors (0% versus 4.5%; P ؍ 0.0036).
The introduction of matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) in the clinical microbiology laboratories is changing the approach to bacterial and fungal identification (1-4). In particular, several studies have already demonstrated the reliability of MALDI-TOF MS in the rapid identification of yeasts in different clinical settings (5-7), evidencing its cost-effectiveness in allowing the initiation of species-targeted antifungal therapy (7-9). To date, four MALDI-TOF MS systems are commercially available: the Microflex LT Biotyper (Bruker Daltonics, Bremen, Germany) (BMS), the Saramis system (bio-Mérieux, Marcy l'Etoile, France), the Vitek MS system (bioMérieux, Marcy l'Etoile, France) (VMS), and, very recently, the Andromas system (Andromas, Paris, France). Several comparative studies have already been performed using the most common systems (BMS and VMS), but, to the best of our knowledge, they have focused only on the identification of bacteria (10-13). Only very recently was a comparative study on yeasts performed using BMS and Saramis (bio-Mérieux, Marcy l'Etoile, France), the previously distributed version of VMS (14). In the present study, we evaluated the ability of BMS and VMS to identify a broad panel of yeasts of medical interest.One hundred ninety-seven isolates from different human samples, previously identified by conventional biochemical techniques or by sequencing the internal transcribed spacer 1 (ITS1) and ITS2 regions, were blindly identified using the two systems. In order to minimize the risk of misidentification related to the use of incomplete and error-filled public databases (15), the sequences obtained were compared to reference data available in two databases: GenBank, searched by using the nucleotide BLAST tool (blast.ncbi.nlm.nih.gov), and the CBS (Centraalbureau voor Schimmelcultures) yeast database (www.cbs.knaw.nl). The panel included 157 (79.7%) isolates belonging to 30 Candida or Candida-related species (Table 1), and 40 (20.3%) isolates belonging to 15 non-Candida species (Table 2). Before processing for MS identification, each isolate was cultured on Sabouraud dextrose (Kima, Padua, Italy) agar and incubated for 24 h at 35°C. For BMS, proteins were extracted as recommended by the manufacturer. Briefly, a loopful of yeasts was suspended in one volume of water and three volumes of absolute ethanol, and after centrifugation, the pellets were processed with an equal amount of formic acid and acetonitrile ...